Risk...measurable, or not?

William McKibbin is running a Linkedin poll asking participants whether or not they think one can measure risk. The response so far suggests that two-thirds feel it is; that's how I voted.

If it's not measurable, does that mean that it's not yet, or that the voters think that the measures are inadequate, or that one can never measure risk, do don't bother trying?

Clearly, we have lots measures that purport to measure risk (standard deviation, beta, downside deviation, value at risk), though we may take exception as to their adequacy.

If you have a moment, cast your vote!